Purpose: In this paper we report outcomes from a group-randomized trial of a prevention curriculum aimed at preventing and reducing bullying and other aggressive behaviors among elementary students in the Denver, Colorado school system. Fourteen elementary schools in Denver were randomly assigned to receive selected modules of Youth Matters, a skills-training curriculum that targets bullying and aggression. Fourteen comparable schools were randomly assigned to a no-treatment control group. Intervention occurred with 603 students in 39 experimental-group classrooms between 2003 and 2005. Four hundred and sixty-two students in 28 classrooms received routine educational services in the control group. Compared to the control group, the experimental group had a slightly, though not significantly, higher proportion of Latino students (64% vs. 49%) and a slightly lower proportion of Anglo (9% vs. 11%) and African American (14% vs. 17%) students. Youth Matters is based on the social development model and incorporates knowledge of risk and protective factors associated with the onset of aggression and other antisocial behaviors in lesson design. Curriculum content teaches students cognitive, behavioral, and social skills necessary to recognize, avoid, and respond to bullying situations in school. The intervention was taught by three curriculum specialists in the 14 experimental schools. Methods: Longitudinal data on self-report measures of bullying behavior were obtained from students over three time points to assess the effects of the intervention on the rate of change in study outcomes. Primary outcomes included a measure of bullying behavior from the Revised Olweus Bully/Victim Questionnaire. A bullying victimization scale and attitudes toward bullying scale also were examined as primary outcomes. Risk and protective factors for involvement in aggressive behavior including measures of school commitment, classroom friction, family attachment, peer influence, and antisocial attitudes were evaluated. Linear growth models were fit to test the effect of the Youth Matters intervention on the rate of change in self-reported bullying behaviors, bully victimization, and other outcomes. Results: A steady age decline in the mean bully victimization score of approximately 3% per semester was observed in the average control group school over the course of study. This age decline is consistent with results reported in prior studies of bullying behavior in elementary school. However, self-reported bully victimization declined at a significantly faster rate over the course of the study in the schools that received the Youth Matters curriculum. Results from linear growth models indicated that participation in the Youth Matters curriculum was associated with a 7% decline in self-reported bully victimization per semester. Over the course of three semesters this intervention effect translates into a 21% decline in the mean bully victim score in the average Youth Matters school compared to only 9% in the average control group school. Implications for Practice: Our findings suggest that increasing skills in recognizing and resisting bullying behaviors reduces bullying victimization among fourth and fifth grade students. Implications of this and other results for school-based prevention strategies are discussed.